Oracle lines up vast funding push for AI cloud expansion

Oracle is preparing to raise about $50 billion during 2026 as it accelerates spending on data centres, specialised chips and network capacity to meet a sharp rise in demand for artificial-intelligence workloads, according to people familiar with the company’s plans. The proposed capital raise, which would rank among the largest ever attempted by a US technology group in a single year, reflects the scale of investment required to compete in an AI-driven cloud market dominated by a handful of global players.

The software and database group has told investors it expects capital expenditure to remain elevated well into its next fiscal year as it builds out AI-ready cloud regions across North America, Europe and parts of Asia. Executives have signalled that existing cash flow will not be sufficient to cover the pace of expansion, prompting the company to explore a mix of debt issuance and structured financing tied to long-term customer contracts.

Demand for Oracle’s infrastructure has been fuelled by a wave of large enterprise and AI-native clients seeking alternatives to the largest hyperscalers. Among those running significant workloads on Oracle Cloud Infrastructure are OpenAI, Meta, Nvidia, AMD, TikTok and xAI, people close to the matter said. These customers require vast clusters of high-performance GPUs, low-latency networking and long-term power availability, driving costs well beyond those of traditional enterprise cloud services.

Oracle’s strategy hinges on locking in multiyear contracts that provide predictable revenue streams against which it can raise financing. Management has argued that AI infrastructure differs from earlier cloud cycles because customers are willing to commit capacity years in advance, reducing utilisation risk. That approach has allowed the company to announce several large AI-focused deals over the past year, some running into the tens of billions of dollars over their lifetime.

Within the company, the funding push is being framed as a necessary response to structural shifts in computing. Training and running advanced AI models demands enormous upfront investment in hardware and energy, while margins depend on operating those assets at near-full capacity. By contrast, Oracle’s legacy database and applications businesses require far less capital but are growing more slowly. Executives have told analysts that the balance of the group is tilting decisively towards infrastructure as AI adoption spreads across industries.

Raising $50 billion in a single year would be a stretch even for a company of Oracle’s size, and market conditions will be critical. Higher interest rates have made large debt deals more expensive, while investors are increasingly scrutinising the long-term profitability of AI spending across the sector. Oracle’s relatively strong balance sheet and history of generating steady cash flow are expected to help, as is its practice of aligning borrowing with contracted revenue.

Analysts note that Oracle is positioning itself as a neutral provider at a time when some customers are wary of relying too heavily on cloud platforms run by direct competitors. That pitch has resonated with AI developers and media platforms that want scale without ceding strategic control. Oracle has also benefited from its willingness to customise infrastructure and pricing for specific workloads, rather than offering a one-size-fits-all service.

Competition, however, remains intense. Larger rivals continue to outspend Oracle in absolute terms, and securing enough power and chips has become a bottleneck for the entire industry. Supply constraints around advanced GPUs, along with regulatory and environmental hurdles tied to new data-centre construction, add uncertainty to expansion plans. Oracle has sought to mitigate these risks by pre-ordering hardware and partnering with energy providers, but execution will be closely watched.